Currency manipulation and the NAFTA renegotiation: with manipulation in remission, now is the time to add currency provisions to the agreement.

Bergsten, C. Fred

U.S. Trade Representative Robert Lighthizer has indicated that the Trump administration will seek to include the currency issue in a renegotiated NAFTA (and in any other trade agreements it might pursue). U.S. Secretary of Commerce Wilbur Ross has reiterated the Administration's focus on bilateral trade balances on many occasions, and has specifically called for appraisal of the impact of currency misalignments, "whether or not due to manipulation," in assessing the causes of U.S. deficits.

The Trump administration would be misguided to try to use trade agreements to reduce the overall U.S. external deficit, which is a macroeconomic problem that requires macroeconomic solutions. It would be doubly misguided to focus on bilateral trade imbalances, whose reduction would do little or nothing to correct the global U.S. deficit, especially with countries such as Mexico and Canada that are running large global deficits themselves. But currency manipulation is an unfair trade practice that can have huge effects on trade flows and trade balances, and it is thus quite appropriate for the administration to address it in their trade negotiations.

A currency chapter in the new NAFTA and other trade agreements could accomplish four things. It should commit the members to avoid manipulation or competitive depreciation. It should encompass a simple definition of manipulation to identify violators of that commitment as incorporated in currency legislation passed by Congress a year ago: substantial intervention in the currency markets by countries with sizable surpluses to block or limit significant appreciation of their exchange rates. It should subject these commitments to the dispute settlement mechanism of the overall trade agreement. Most critically, it should provide effective sanctions against violators in order to deter future manipulation, including a "snapback" of the trade advantages that the violators had gained under the agreement.

Currency manipulation is the use of intervention in the foreign exchange markets by countries to avoid or limit appreciation of their exchange rates. Largely because of massive intervention by China during 2003-2013, along with more isolated but sometimes sizable intervention by Japan, manipulation came to be viewed as an unfair trade practice and was widely criticized as such throughout the Congressional debate over trade policy during the current decade and in the political campaigns in 2016. In my new book with Joseph Gagnon, Currency Conflict and Trade Policy, we analyze the issue in depth and conclude that it was quite significant: about twenty countries were active during the "decade of manipulation," their excessive intervention averaged over $600 billion per year, and the result was a shift of more than $300 billion of annual current account balances. The U.S. current account deficit increased by $150-$200 billion annually as a result and the United States lost more than a million jobs during the Great Recession and the tepid recovery from it.

A large number of members of Congress rejected the responses of both the George W. Bush and Obama administrations to the problem as grossly inadequate, and urged the inclusion of "enforceable currency disciplines" in the Trans-Pacific Partnership. Congress included the issue as a major U.S. negotiating objective in the Trade Promotion Authority legislation in 2015 and passed new currency legislation as part of the Trade Enforcement and Trade Facilitation Act of 2015 (the "customs bill") that adopted objective criteria for identifying manipulation (see below) and required presidential response under some circumstances. The Obama administration, in an effort to deflect Congressional pressure without jeopardizing the TPP, negotiated a side agreement to it that reiterated the commitments of the member countries (from the IMF Articles of Agreement) to avoid manipulation and to enhance the transparency of their related international financial activities, but included no dispute settlement mechanism or enforcement provisions. It also created a TPP Macroeconomic Group that was somewhat analogous to the consultative North American Financial Group that was created at the outset of NAFTA but was not active for very long. But many members of Congress remained dissatisfied with the policy response and the issue was thus cited widely, in the Congressional debate on TPP and the election campaigns in 2016, as a reason to oppose any new trade agreements and indeed globalization more broadly.

Both Canada and Mexico maintain floating exchange rates that are largely "clean," that is, conducted without intervention. Neither has been accused of manipulation in recent decades. Mexico in fact recently sold modest amounts of dollars to counter depreciation of the peso.

Furthermore, unlike China or Japan, both Canada and Mexico now run sizable global current account deficits (despite Mexico's bilateral surplus with the United States) on the order of 2.5 percent to 3 percent of their GDPs. For both these reasons, the practical impact of adding currency issues to NAFTA would be modest or even non-existent at present. However, the combination of domestic political pressure in the United States and the Administration's desire to set a precedent for other trade agreements where currency might be more relevant (for example, Japan, a revived TPP) suggest such an effort. The blueprint suggested below outlines how such a chapter could be fashioned and incorporated in a revised NAFTA (or bilateral agreement with Mexico).

HISTORICAL BACKGROUND

Trade agreements have traditionally avoided trade imbalances and the currency issue for two reasons. First, such agreements aim primarily at expanding the level of trade, and sometimes at affecting its structure and the rules governing it, which are microeconomic issues. By contrast, trade balances are primarily a reflection of saving-investment differences and broader economic fundamentals that can be addressed effectively only through monetary, fiscal, and other macroeconomic policies including exchange rates. In addition, trade agreements are explicitly or implicitly premised on the principle of reciprocity in reducing tariffs, subsidies, and other distortions; because of domestic politics in all the countries involved, they are not intended to adjust trade imbalances.

Timing has also played a major role in this traditional differentiation. Trade imbalances, and the currency misalignments that can produce them, have been seen as transitory developments that will self-correct (by markets under flexible exchange rates) or be corrected (by governmental policies under fixed exchange rates) within relatively short periods. The guidelines for IMF surveillance of countries' exchange rate policies inveigh against "protracted" intervention in the currency markets because it can prolong imbalances beyond their normal short-run horizons. Trade agreements, by contrast, are intended to permanently alter economic relations between participating countries.

The second, institutional, issue is that currency policy and trade policy are generally managed by different authorities. Finance ministries and central banks (which are often independent from governments) are usually responsible for exchange rates, and trade or commerce or foreign ministries handle trade policy. At the international level, the International Monetary Fund is responsible for exchange rates and the World Trade Organization covers trade. Turf conflicts between these actors have frequently prevented a coordinated response to issues that link currency and trade, both within and among countries.

The United States has tried to coordinate trade policy and international monetary policy on occasion over the years. In 1971, President Nixon imposed a temporary import surcharge to help negotiate devaluation of the overvalued dollar, a step partly taken under pressure from Congress. In 1985, Secretary of the Treasury James A. Baker III, responding even more centrally to congressional anxieties about trade deficits, negotiated the Plaza Accord to weaken the overvalued dollar and strengthen the European currencies and Japanese yen.

Lack of coordination is much more common, however. The IMF and GATTAVTO have frequently discussed better coordination and set up mechanisms to promote it but without much success. The IMF staff vetoed inclusion of currency considerations in China's protocol of accession to the World Trade Organization on the grounds that such a provision was within the jurisdiction of the Fund rather than the World Trade Organization, an under-the-radar decision that had profound effects in light of the major role that currency manipulation played in the subsequent explosion of Chinese exports and trade surpluses.

These substantive and institutional considerations have traditionally led the manipulation issue, like all currency issues, to be addressed by monetary officials and the International Monetary Fund rather than by trade officials and trade agreements, including the World Trade Organization. The United States attempted to pursue this approach, primarily with respect to China, for most of the past decade. Problems can arise when trade imbalances persist, however, particularly if they do so at least in part due to demonstrably unfair trade practices that violate the agreed international rules. The United States, for example, has run sizable current account deficits for almost all of the past thirty-five years. China has run sizable surpluses since the early 2000s, most or all of which throughout the "decade of manipulation" could be explained by its currency practices.

Hence the monetary efforts lost much of their credibility, and U.S. Treasury Secretary Jacob Lew acknowledged in 2015 that a trade agreement has to be built on firm commitment to market-determined exchange rates. This "monetary issue" played a major role in the widespread rejection of the Trans-Pacific Partnership and opposition to other trade agreements, including NAFTA, and indeed to globalization more broadly, that pose major risks to the openness of the global trading system. There is thus a strong case for incorporating manipulation in future trade agreements, as advocated widely in Congress and now by the Trump administration.

A BLUEPRINT

Most trade agreements, including those negotiated by the United States, include chapters on specific topics. A currency chapter (or parallel side agreement) in a renegotiated NAFTA would ideally include four components: a statement of objectives, criteria for defining and pursuing those objectives, a decision-making process to implement the agreement, and policy responses to enforce their implementation. These elements should, where possible, conform to existing international agreements, notably the Articles of Agreement of the IMF and the charter of the WTO.

The objectives of the NAFTA currency chapter could be drawn from the IMF Articles, as are the objectives of both the currency chapter of the Trade Facilitation and Trade Enforcement Act of 2015 and the Joint Declaration of the TPP Monetary Authorities. IMF members have accepted obligations to "avoid manipulating the exchange rate or the international monetary system in order to prevent effective balance-of-payments adjustment or to gain unfair competitive advantage over other members" (Article IV, Section 1 (iii)). The Fund is supposed to maintain surveillance over exchange rate policies and discuss "protracted large-scale intervention in one direction in the exchange markets" with errant members. The Articles also call on member countries to "take into account in their intervention policies the interests of other members, including those of the countries in whose currencies they intervene." These precepts could provide the foundation for specifying the goals of a currency component of a free trade agreement and could even be incorporated simply by reference; the free trade agreement would then seek to provide an enforcement mechanism, which is lacking in the IMF itself.

The G-7 (Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States) has also adopted a commitment to consult within the group before undertaking intervention activities, and members have largely adhered to that agreement. G-20 communiques have pledged that members "will not target our exchange rates for competitive purposes," though some of them have ignored that stricture (denying that their targeting is done for competitive purposes). Such nonbinding pledges could be incorporated into trade agreements if binding commitments were not possible.

The methodology for pursuing the agreed objectives could start with commitments to provide data on the relevant variables, per agreed IMF conventions in most cases, and as in the TPP Declaration. These commitments should include reserve levels, including those outside official monetary reserves (notably in sovereign wealth funds); intervention; the currency composition of official reserves; and currencies of intervention.

Determining the existence and extent of currency misalignment, especially as a possible trigger for remedial action, has proven enormously difficult both intellectually and politically. Numerous conceptual approaches to defining and measuring currency "misalignment" have been attempted. The IMF uses three different measures, which often produce very different results. Most official discussions, and even many academic efforts, have foundered at this initial level.

Trade agreements like NAFTA should thus ignore the determination of "misalignment" per se in favor of objective indicators, as does the Trade Facilitation and Trade Enforcement Act of 2015 (drawing on the initial Bergsten and Gagnon study on this topic in 2012). The goal would be to deter a country from running large and persistent external surpluses that result from depressing the value of its exchange rate in the currency markets. Only three variables need to be identified: the level of reserves (to determine if they are "excessive"), the size of the current account surpluses, and the extent of intervention (or changes in reserve levels as a proxy if actual intervention figures are not available on a timely basis).

The Treasury Department has to date interpreted the new law as indicting major trading partners of the United States, defined as the twelve largest of those partners (of course including Canada and Mexico), that run current account surpluses exceeding 3 percent of their GDPs and conducting intervention in excess of 2 percent of their GDPs over the preceding twelve months (as well as running bilateral surpluses of more than $20 billion with the United States). No country now meets all three criteria, although six meet two of them and have thus been placed on a "monitoring list": China, Germany, Japan, Korea, Switzerland, and Taiwan. The Treasury reports have not addressed Canada or Mexico because they have not been relevant to the issue of manipulation.

A key concept is "intervention." Substantial direct purchases of foreign exchange with domestic currency should be a central criterion for triggering a currency provision in a free trade agreement. Participating countries should fully disclose their intervention activities, though some reporting could remain confidential if necessary.

More complex questions surround "oral" and indirect intervention. Oral intervention--that is, calls for market exchange rates to be adjusted unaccompanied by any new policy--can be obvious or extremely subtle. It can have powerful effects, at least in the short run. If new rules limiting direct intervention are credible, however, oral intervention would be less effective because no policy follow-up would be permissible.

Indirect intervention could include a wide range of policies, such as capital controls on inflows and/or outflows and macroprudential financial regulations (and particularly the timing of their installation and removal). It would be extremely difficult to define such measures with precision, however, because many steps seen as indirect intervention could be defended as having much broader purposes. Macroeconomic policies, including quantitative easing, and fiscal policies, for example, should not be included. All international rules and norms, including those of the IMF, the G-7, and the G-20, explicitly recognize this distinction and exonerate quantitative easing policies from any responsibility for currency manipulation. Some U.S. officials have nevertheless worried that U.S. monetary policy could be attacked under any new international currency rules, and have been reluctant to pursue such rules for that reason.

Particularly in the case of measures with indirect effects on exchange rates, intent can be an important consideration. Were the steps undertaken to influence exchange rates or were such influences solely a byproduct of some other primary purpose (as in the case of quantitative easing)? The requirement to demonstrate intent to devalue competitively under current IMF doctrine has enabled countries to defend clearly manipulative actions in the knowledge that no mechanisms exist to override their assertions. This problem underscores the need for objective indicators, such as direct intervention and current account surpluses, as triggers for action.

As for the decision-making process through which these concepts could be implemented, traditional practice in both the WTO and existing free trade agreements provides clear guideposts:

* An aggrieved country requests consultation with the alleged violator of the rules, and a major effort is made to reach a mutually satisfactory voluntary solution.

* Failing agreement in the consultations within ninety days (or some other tight time limit), a panel with the relevant expertise is chosen (from a contingent list, with particular expertise on currency issues) to recommend a solution within another tight time limit (another ninety days).

* A country found to have violated the rules and failed to accept the recommended solution within another tight time limit is subject to the penalty phase, in which a separate compliance panel (perhaps comprising the same experts) authorizes countermeasures.

* That panel monitors the situation, taking into account the expected lagged effects of previous exchange rate changes in eliminating the excessive current account surpluses, and calls for termination of retaliation when the cause of the problem (those surpluses or the manipulation) ceases.

The final question is what enforcement mechanisms could be included to make the agreement work and ensure its credibility. The absence of such mechanisms has been a cardinal flaw of the IMF system throughout its existence and a chief source of congressional criticism. Five types of measures are possible: withdrawal of concessions made in the free trade agreement itself, imposition of import surcharges, imposition of countervailing duties, monetary penalties (fines), and countervailing currency intervention. Gradation of each measure is possible, with initially modest penalties subsequently intensified in accordance with the seriousness or extent of the violation.

The usual technique for withdrawing concessions in a free trade agreement is the "snapback clause," under which tariffs are returned to the pre-agreement level (usually the most-favored-nation rate) for "breach of the agreement." Snapbacks are typically applied on a product-specific basis, to counter violations in a particular sector, but they could be installed across the board in the case of currency violations. It would also be possible to apply the snapback concept to concessions other than tariffs, as in a recent WTO case in which Brazil was authorized to withdraw some of its commitments regarding intellectual property rights if the United States continued to violate the dispute settlement panel's ruling on its cotton subsidies. The original concessions would be restored when the problem was corrected.

More extensive retaliation can be envisaged, including the imposition of import surcharges or countervailing duties, if currency manipulation is deemed a countervailable subsidy like any other. Monetary penalties--like those NAFTA imposes for violation of its labor and environmental disciplines and through which the United States has compensated Brazil for U.S. violation of WTO agricultural agreements in the cotton sector--could be added to the arsenal of potential measures. A key problem with each of these options other than snapbacks, however, is the difficulty of calculating the amount of the currency undervaluation to provide a basis for determining the magnitude of the permitted retaliation. Because such calculations are fraught with both intellectual uncertainty and political sensitivity, as noted above, they should be avoided. The snapback approach should be the chief trade policy response to manipulation under an free trade agreement.

This also means, however, that it would be desirable to add a monetary policy tool that would fight fire with fire. Such an approach would also overcome the problem that trade policy remedies, like snapbacks, curb only imports whereas currency manipulation also suppresses the aggrieved country's exports (to global markets as well as to the manipulating country itself). An aggrieved country could be authorized to purchase the currency of the manipulating country in the amounts needed to neutralize the impact of that country's own intervention in the foreign exchange markets upon a finding by the dispute settlement panel that manipulation was taking place. A clear indication by the United States that it was prepared to act on such authorizations should deter most if not all future manipulation efforts.

The U.S. Treasury and the Federal Reserve could carry out such "countervailing currency intervention" under current legislative authorities. Specific authorization for such a policy was nevertheless included (as "remedial currency intervention") in a currency bill passed by the Senate (but not taken up by the House) in 2011, and it would be desirable for the Congress to enact such legislation now, including to authorize Treasury to borrow additional amounts if needed to fund any actual CCI purchases. (These would be no budget cost, because the purchase of foreign currencies would represent an "exchange of assets" with the dollars sold, and the policy should in fact make money for the U.S. taxpayer because the purchases would by definition take place when the foreign currencies were substantially undervalued). Including such a provision in free trade agreements would be a straightforward and effective deterrent to currency manipulation by parties to the agreements. Lodging implementation of this key sanction in finance ministries and central banks should assuage most of the institutional concerns that normally make it difficult to address the problem through trade agreements.

Mexico can be expected to resist inclusion of a currency chapter in a renegotiated NAFTA, as it did when the idea was raised informally during the TPP negotiations (although it did sign the TPP side agreement that was eventually worked out, presumably because that agreement carried no new obligations for them). It might thus prove impossible to agree on binding and comprehensive rules subject to an effective dispute settlement mechanism and consequent sanctions. Compromises might be needed on such matters as the ambition of the rules, the degree to which they become legal commitments, the rigor of the dispute settlement mechanism, and the severity of the sanctions against offenders. Tradeoffs among such variables are likely in any NAFTA renegotiation. (On the other hand, Canada under its previous government was reportedly one of the few TPP countries that was receptive to U.S. entreaties on the issue.)

The most plausible wiggle room lies in the ambitiousness of the criteria that would trigger action. The term "excessive," as applied to levels of reserves and intervention as well as current account surpluses, could be set high enough that only the most egregious violators would be caught, which is in fact the basic objective of the exercise.

Another possible avenue of compromise relates to the interaction of the obligations binding the participants and the methodology through which they are implemented. Countries wishing to limit their risk of exposure will want to trade off "soft" obligations against "hard" dispute settlement provisions or vice versa. For example, the indicators of violations of the agreed currency obligations could become "presumptions" or even "illustrations," including in side agreements as in the TPP, rather than legally binding commitments. The adjudicatory panels could be limited to recommendations to a politically constructed final arbiter rather than binding protocols, as is the case with dispute settlement mechanisms in some existing free trade agreements. It would be perfectly plausible to set up and finely tune a separate dispute settlement mechanism for the currency chapter (or side agreement) as part of the overall negotiation of the issue.

The inclusion of currency provisions in a renegotiated NAFTA, or any new bilateral U.S. trade agreements (including with Mexico), would represent a major change in the structure of such compacts. There might turn out to be a parallel with the U.S. insistence on including labor and environmental provisions in its free trade agreements over the past two decades: those additions were strongly resisted by other countries (and most proponents of free trade) at first but have gradually become an accepted component, in response to U.S. domestic political pressures but increasingly because their merits have become widely (if grudgingly) accepted. The trade-currency linkage, which is substantively indisputable, could follow a similar pattern and a modernized NAFTA could be an initial step along that path--as the original NAFTA was an initial major step toward the widespread inclusion of labor and environmental issues themselves in trade agreements.

Trade and currency policies have traditionally been conducted separately by the United States and most other countries. That bifurcation is no longer viable in the United States, however, as demonstrated by the central role played by concerns over currency manipulation, based on its very substantial costs to the U.S. economy, in the backlash against trade agreements and globalization more broadly.

Manipulation is now largely in remission but much more forceful policy responses to it will almost certainly be a necessary component of any sustainable new political foundation for an open foreign economic policy in the United States. One part of that response could be inclusion of effective provisions to deter future manipulation in new U.S. trade agreements, starting with the renegotiation of NAFTA. There would be little practical effect from such a step, since neither Canada nor Mexico has been accused of currency manipulation in recent decades and both are in fact running large global current account deficits (despite Mexico's bilateral surplus with the United States). This should make inclusion of such provisions less controversial and more feasible, which will probably be necessary for the agreement to win congressional approval and would set a useful precedent for future agreements with countries (such as Japan and other former TPP partners) where the issue would be more salient.

BY C. FRED BERGSTEN

C. Fred Bergsten is Senior Fellow and Director Emeritus of the Peterson Institute for International Economics.